Approximate reasoning for production planning (Q1112714)

In one part of the paper the author cites in a brief form a series of his earlier results, which concern the so-called interval-valued fuzzy sets (IVFS's). An IVFS is obtained as the result of a generalized operation S(A*B) performed on usual point-valued fuzzy sets A and B (``*'' denotes one of the operators ``and'', ``or'', ``\(\to '')\). This operation is defined as follows: \[ (1)\quad S(A*B)=\{s| \quad DNF(A*B)\subseteq S\subseteq CNF(A*B)\} \] where DNF and CNF are the disjunctive and conjunctive normal forms, respectively, of a fuzzy composition being an extension of the canonical forms in Boolean logic. For example, for the operator ``and'' DNF and CNF are defined as follows: \[ (2)\quad DNF(A\quad and\quad B)=A\cap B\quad and\quad CNF(A\quad and\quad B)=(A\cup B)\cap (A\cup B^ c)\cap (A^ c\cup B) \] where \(\cup\), \(\cap\) and superscript c are used in the set notation to correspond to the intersection, union and complementation operator, respectively. However, they have to be such operators that the containment condition is satisfied: DNF(A*B)\(\subseteq CNF(A*B)\). The author presents two general classes of t-norms and t-conorms, which have this containment property. By using various t-norms and corresponding to them t-conorms belonging to these classes, many new representations and interpretations of linguistic compositions may be obtained. In the following part of the paper the author proposes a generalization of the operation (1) for the case when A and B are not point-valued, but interval-valued fuzzy sets. The author suggests to use the generalized operation S(A*B) in constructing rules of approximate reasoning. In the paper the author presents an example of the application of such rules in a problem of production planning.